I have a tube of X23-7921-5. I thought the 7783D wasn't good for laptops?
Sent from my HTC One_M8 on Tapatalk
-
Wasn't good for me, but it may have something to do with me not burning it in properly (apparently it needs to be burned in for 30 minutes at >60C).
-
Honestly I didn't know it had a burn in period. What info is I usually game pretty hard after repasting for at least 1-2 hours and most likely unknowingly burned it in.
-
Better than TX-2, NT-H1 and AS5 probably as well. None of those pastes have been durable for me in highly stressful conditions.
-
Noctua NT-H1 thought I was a terrible disappointment. I can not understand that this thermal grease has received so much praise. Most cooling pastes work well on a desktop pc since it has better cooling and can maintain a lower temperature. High temperature = rapid destruction of thermal grease
-
Yeah Noctua NT-H1's pumping out effect has been documented (read the last 2 lines in that linked post).
On a broader note, the original thread itself is a good read, since de-lidded bare die mount now mimics the situation in laptops. In a way it's also complementary since desktop heatsinks easily have twice the mounting pressure compared to laptops. I imagine the pastes that suffer from pumping out would fare even worse under high pressure situations.
Just don't take the Liquid Ultra results too seriously, he seriously applied way, way, WAY too much Liquid Ultra, seriously.
(also in this post you can see the scratches left by IC Diamond, perhaps caused by improper removal)D2 Ultima likes this. -
The difference between the Liquid Ultra and Noctua nt-h1 was over 10 degrees celsius on my processor
-
Yeah I remember seeing something like a 7C drop as well. I think the poor results IDontCare got was due to using too much Liquid Ultra.
-
Wow, I'm getting some! 7-10C drop is significant. -
Liquid ultra was worthless on my CPU because the contact pressure doesn't exist period. The paste completes the connection to the heatsink :/
Sent from my HTC One_M8 on Tapatalk -
Clevo is still having little to no pressure on their heatsinks, huh? That's unfortunate
-
My first post back here after almost 3 weeks of being gone and I just have to laugh XD.
-
Yeah seems to be hit or miss. I got really lucky with my CPU heatsink though as the contact is great except for 2 corners. Probably why I saw a huge improvement with Liquid Ultra.
-
I'm excited as hell to try liquid ultra. I think my heatsink has a decent contact because when the MX-2 pumped out, I was still able to remain under 90 degrees doing almost every application I could think of (except BF4 in the daytime)... and that was with basically no thermal paste. So... yeah. I would kill for some liquid ultra when I get back my PC. Please note: I don't have it back yet. T-T
-
OT but related to D2 Ultima.
I found this in your SLI guide at LTT Forums under the first spoiler:
This is incorrect. Multiple GPU's in SLI can function just fine with asynchronous clocks and the faster card will not downclock itself to match the slower card. This is why I can overclock my two cards asynchronously in Nvidia Inspector and see performance improvement compared to if both cards were running the lower clocks. And with the advent of GPU Boost, it's the norm for otherwise identical cards to run at different clock speeds in SLI by default since max Boost clocks are dependent on each individual GPU's ASIC quality. -
Well, I currently have the GELID Extreme paste (actually Phobya Extreme but they both use the same OEM) on the most extreme situation possible, my Sapphire Tri X 290 overclocked to 1200mhz with +165mV with relatively poor contact (though good pressure). It's been 3 months so far and no issues, MX-4, Phobya standard, Arctic Silver all couldn't take it (they all had crazy pump out issues after a a few hours or days). ICD was fairly stable but the conductivity couldn't cut it.Ethrem likes this.
-
Poor contact but good pressure??? Do you mean the areas that do have contact have good pressure, but contact on the whole is lacking?
-
Well good hopefully I don't have to repaste my CPU for some time then.
Sent from my HTC One_M8 on Tapatalk -
The contact surface area between the heatplate and the die was not as intimate as I would like, however, the mounting mechanism is extremely tight and allowed very heavy pressure so there's no buckling or movement of the assembly.
-
Someone else was informing me that this happened on those same forums. I didn't have my computer to test it, but back when I wrote the guide I saw no improvement in performance unless I boosted both cards. I guess they must have changed it, but since you're the second person to tell me this, I'll have to update it.
Edit: Why in the living daylights can I not give rep to people? Have I really only given it to approximately 4 different people in the last like 6 months and now I have to give to more? O_Oocticeps likes this. -
How about try Cool Laboratory Liquid MetalPad? CoollaboratoryLiquid MetalPad - the innovation of cooling for High-End sytems Place two pieces on top of each other
-
Yes XD
If it's anything like Indigo Xtreme then NO JUST NO HELL NO KILL IT WITH FIRE.D2 Ultima likes this. -
Wow those are some awful photos. Heatsink is completely out of focus. All I can see are the heat fins on top of his blingy RAM.
-
What about copper pad shims? Adding one piece on top of the CPU die to increase pressure :hi2:
-
That sounds like asking for a cracked die...
-
Careful when you attach the heatsink :laugh:
-
Yeah my 4940MX already looks like a disaster area thanks to ICD scratching the heck out of it, I don't want to risk actually cracking the die.
I just worked around it. It has GC Extreme paste and I drop my TDP to 47W with an 85mv undervolt and it manages my temperatures while still giving me 3.5-3.6GHz under full load and 4GHz with one core, 3.9 with two.
Sent from my HTC One_M8 on TapatalkBaoTCP likes this. -
I still say, get a closed-ear headset and tell your roommate to steam-message you or yell when he needs you, then put your fans on max =D
-
I'm glad I'm not the only one who dislikes ICD...used it on my 2820QM a few years ago and it tore it up! Went from mirrored finish to scratched haziness....
-
ICD doesn't scratch CPU dies period. In all of our completely fair and independent testing we were never able to replicate these so-called "scratches". Stop propagating this slanderous myth. [/IC Diamond]
-
It actually looks so bad I thought the die was cracked.
Was not a happy camper!!!
Sent from my HTC One_M8 on Tapatalk -
Umm... i use icd. It scratches the hell out of dies.
-
As much as dislike the scratches on my die, ICD is still the best paste for me due to its longevity under heavy load as well as it's top performance. the fact that it is really thick makes it better (for me at least) on lower contact pressure situations where Liquid Ultra would not be able to make adequate contact at all. I find the benefits outweigh the drawbacks, most of the time.
Plus I'm not using the dies as a mirror to put on make up for a drag night out. The dies sit under a heatsink and the paste bridges that gap, something ICD does very well.Marksman30k likes this. -
What other pastes have you tried besides ICD and Liquid Ultra?
-
Which paste should I try next? Been using OCZ Freeze for a few years and apparently it's not very good, at least according to these charts LOL:
Charts, benchmarks Thermal Compound Charts, [5] CPU Air Cooling, Low Pressure
Charts, benchmarks Thermal Compound Charts, [6] GPU Cooling
So far the top candidates seem to be Gelid GC Extreme and Prolimatech PK-3. Neither Shin-Etsu X23-7921-5 nor Xigmatek Xi-3 HDT look to have good availability. In the case of the latter, I can't find it online at all.
Any other suggestions or recommendations? Same requirements as OP--no liquid metal or ICD.
-
Moddiy has Shin Etsu. It's not badly priced either.
http://www.moddiy.com/products/Shin...al-X%2d23%2d7921%2d5-W{47}m.K->-6.0-(3g).html
Can't beat 14.99 and free shipping from Hong Kong
With that said, GC Extreme has had the best performance for me so far (I haven't tried the Shin Etsu as of yet)
Sent from my HTC One_M8 on Tapatalk -
Don't know if you've seen this thread yet. It's invaluable since it's tested on a de-lidded 3770K with a bare die mount, so it's about as close to a laptop situation as you can get. Noctua NT-H1 had great results but as the author said himself, it has pumping out issues. Arctic Ceramique seems to be the second best after that.
I can also vouch for Gelid GC Extreme, at least for the 2 weeks I used it before switching to the ShinEtsu. Still the most viscous paste of all the ones I tested, so should really help with low pressure, poor contact situations.Ethrem likes this. -
That test is rather interesting indeed. I might have to invest in some Noctua. I refuse to buy any more IC Diamond.
How are your temps holding with the Shin Etsu?
Sent from my HTC One_M8 on Tapatalk -
Nah don't bother with Noctua, it's got pumping out issues which means it will epic fail within a month if even that.
Temps holding just fine with ShinEtsu, still within 1-2C either direction of what I had in my first post.Ethrem likes this. -
My temps are holding with GC Extreme too. I really wish I had better contact with my heatsink though, my 4940MX has so much potential.
Sent from my HTC One_M8 on Tapatalk -
Yeah that's the only place which stocks the Shin-Etsu. Both PK-3 and GC Extreme are cheaper and have better listed conductivity though.
-
It's much more complicated than a simple thermal conductivity, there's also the thermal resistance of the TIM itself, as well as factors like bond line thickness (BLT) you have to consider. Check out the equation on page 10 of this PDF, see the variables go into that single "thermal resistance" component?
Also, I linked to a paper a few pages back where it basically showed that for solder TIMs at least, if you can get the bond line thickness down to 0.1mm, then by that point conductivity is no longer a bottleneck in thermal transfer. Not surprising given the equation above, and also explains why a higher mounting pressure is preferred -- high pressure = lower BLT = lower overall thermal resistance -
I've said this before: Noctua is a disappointment. Use Gelid extreme it is a better thermal grease. I think it's the best pasta except Liquid Ultra
-
Expensive, but this stuff looks really good too based on the few reviews I've seen:
JunPus Nano Diamond Thermal Grease JP-DX1 ,Overclocking Thermal Compound , Best Thermal paste (3g) - Newegg.com
Junpus JP-D9000 Nano Diamond Thermal Grease /Overclocking Thermal Compound / LN2 Thermal paste (3g) - Newegg.com
Only thing is it seems to have similar composition to ICD, so there is that worry about scratched dies. -
I'd personally avoid anything with "Diamond" in its name
Seriously though, do you care more about longevity or short term happiness? -
What about Phobya NanoGrease Extreme? I know that Phobya HeGrease Extreme (ROFL sounds like what a bodybuilder does) is just a rebranded Gelid GC Extreme, but is NanoGrease better?
So confusing. I should just buy the cheapest "good" paste and call it a day. :laugh: -
I hear ya.
It's gotta be a good performer and consistent over a long period of time. I don't want something that gets amazing temps at first but needs to be reapplied often. I'd rather have something that's a little less amazing but lasts. At the other end of the spectrum, ICD and Liquid Ultra are out of the picture. -
I have a tube of NanoGrease, it seems slightly better than HeGrease but it is really hard to apply on account of the fact it has really poor surface wetting. Sometimes I feel I need to use some kind of surface tension remover to get this to integrate to the surface. Costs an arm and a leg though, I never restocked, just got me the last 5 tubes of Phobya HeGrease Extreme in the store and called it a day.
Depends on how much of a PITA the laptop is to repaste. If it's a Dell, HP, Acer or Toshiba then definitely ICD. Not much point using ICD for Clevo or Alienware machines unless the HSF contact is THAT bad, in which case an RMA of the assembly or a copper shim might do better. -
Screw it. I'm getting Prolimatech PK-3. It's the cheapest "premium" paste.
ShinEtsu thermal paste
Discussion in 'Alienware 18 and M18x' started by n=1, Aug 1, 2014.